Insulin infusions in the neonatal unit: delivery variation due to adsorption

To assess the extent of the variability in insulin delivery over time, under conditions used in Australian neonatal units, studying the following variables: diluent, preconditioning, flushing, sequential preconditioning and flushing, insulin concentration, flow rate, catheter type, and addition of albumin.

A range of simulated insulin infusions was studied. Infusions were run over 22 h, with aliquots of infusate collected at baseline and after 15 min, 30 min, 1 h, 2 h, 6 h and 22 h. Insulin concentration was assayed using a radioimmunoassay.

An infusion of 50 mU insulin/mL at 1 mL/h gave negligible insulin delivery until 22 h. However, after preconditioning and flushing the tubing, consistent insulin delivery was attained by 6 h. An infusion of 200 mU insulin/mL at 1 mL/h did not provide consistent delivery until 6 h. At this concentration and rate, consistent insulin delivery was attained within 30 min of the start of the infusion either by preconditioning and flushing the tubing, or by addition of albumin to the infusate.

Clinically significant variation in intravenous insulin delivery will occur in the neonatal setting unless counter-measures are taken, such as sequential preconditioning and flushing of the delivery system or the addition of albumin to the infusate.